Knowledge of the 3-D structure of proteins is essential for understanding how they function and the designing of drugs to affect those functions. The Structural Genomics Initiative of the NIH aims to solve the 3-D structure of thousands of proteins a year by a major scale up of technology. X-ray crystallography, which requires high-quality protein crystals, is a significant means of deducing 3-D structure. Growing protein crystals is a process of trial and error and, in high-throughput protein crystallization, essential for the Structural Genomics Initiative to succeed, combinatorial methods are used to set up thousands of crystallization vessels in parallel and machine vision robots inspect each vessel for the presence of crystals and crystal-like objects. A bottleneck is the inability of current robots to distinguish between micro crystals and other small, uninteresting objects, necessitating human intervention to make the distinction. In this Phase I SBIR proposal the feasibility of a novel imaging method will be tested. This method appears capable of distinguishing crystalline objects from amorphous ones and protein crystals from uninteresting salt crystals. Once the feasibility is demonstrated, the new method can form the basis of a robot possessing intelligent machine vision. In Phase II a prototype of the robot will be built. There is demand for such a robot among high-throughput laboratories, pharmaceutical companies, contract X-ray crystallography companies and drug discovery firms. The robot also will play an important role in the success of the Structural Genomics Initiative and the discovery of new drugs.